Cosmogenic 3He and 10Be denudation rates in the Central Andes: Comparison with a natural sediment trap over the last 18 ka

Abstract

It is of major importance for Earth surface sciences to reconstruct denudation rates in the most precise and accurate way. For this, it can be useful to test on the same setting methods based on different assumptions, such as those relying on geomorphological and geochemical observations. Here, we use an exceptionally suited setting in the Locumba catchment (southwestern Peruvian Andes) that offers the unique opportunity to compare denudation rates derived from in situ cosmogenic 3He and 10Be with a geomorphological sediment budget integrated over the last 18 ka. The sediment budget is estimated by determining the volume of sediment trapped in the Aricota lake that formed 18 ka ago after the occurrence of a giant rockslide dam. We reconstructed the topography of the Locumba valley before the dam emplacement and established that the captured sediment volume is 0.8 ± 0.1 km3. Considering that the lake-water output is restricted to seepage through the dam and that overflow above the dam never occurred, this volume correctly represents the sediment flux integrated over the last 18 ka. Integrating this volume over the upstream catchment area (∼1500 km2), we derived a corresponding mean erosion rate of 30 ± 9 mm.ka−1. Fluvial sediments feeding the Aricota lake were sampled to derive denudation rates from in-situ cosmogenic 10Be in the silicates and from in-situ cosmogenic 3He in the ferromagnesian minerals. Cosmogenic nuclide denudation rates from the main stream are 30 ± 2, 33 ± 2, 21 ± 1 and 82 ± 5 mm.ka−1 for the 10Be-quartz, the 10Be-feldspar, the 3He-amphibole and 3He-pyroxene, respectively. The consistency between the cosmogenic nuclide denudation rates derived from 10Be in the silicates and the erosion rate derived from our sediment budget shows that the 10Be accurately estimates of the sediment flux. Additionally, this work provides the first successful application of 10Be-feldspar nuclide-mineral pair to derive catchment-mean denudation rate and demonstrate that 10Be-feldspar can thus be a good alternative in catchments dominated by volcanic rocks with no quartz. The discrepancies observed between the denudation rates derived from the 3He-amphibole and 3He-pyroxene couples require further studies.